Using a logical model to predict the growth of yeast.

Area of Science:

• Systems Biology
• Metabolic Modeling
• Computational Biology

Background:

• Utilized iFF708, a Flux Balance Analysis (FBA) model of Saccharomyces cerevisiae metabolism.
• Augmented the FBA model with data from the KEGG pathway database.
• Employed predicate logic for knowledge representation to explicitly define metabolic network structure.

Purpose of the Study:

• To construct a logical model of S. cerevisiae metabolism.
• To leverage predicate logic for enhanced metabolic network representation and inference.
• To improve model identification and refinement capabilities.

Main Methods:

• Developed a logical model by integrating an existing FBA model (iFF708) with KEGG pathway information.
• Represented metabolic network structure using predicate logic.
• Evaluated model performance by comparing its predictions against empirical minimal medium growth data and essential gene lists, using iND750 as a benchmark.

Main Results:

• The logical model incorporates 263 additional putative genes and 247 additional reactions compared to the original FBA model.
• Performance evaluation against empirical data showed the logical model's predictive accuracy.
• Comparison with iND750, an updated FBA model, assessed the logical model's correctness.

Conclusions:

• The logical model, despite its simpler form and expanded coverage, demonstrated no significant performance degradation compared to the iND750 model.
• Receiver Operating Characteristic (ROC) analysis and statistical studies support the model's efficacy.
• Predicate logic offers a viable and effective approach for metabolic modeling, enabling robust inference and model improvement.